Caveat verdict
clawd-throttle
The skill appears to be a legitimate tool for routing LLM requests to different models based on cost and capability. It does not contain any obvious malicious code and seems to follow standard practices for API key handling and logging.
⚠ Flagged for review — coarse, uncorroborated signal, not a confirmed exploit. Review the config yourself before installing.
Automated static analysis — not a human review. Caveat flags capabilities, not confirmed intent, and can produce false positives. Disagree with this verdict? Use Dispute below.
Findings (5)
Possible hardcoded credential
scripts/setup.sh · prose · downgraded · API_KEY: " ANTHROPIC_API_KEY
[ -n
Pipe to sh — executes piped content as shell commands
src/classifier/dimensions.ts · prose · downgraded · |sh
Downloads executables from external URLs
scripts/setup.sh · prose · downgraded · Install from https://
Dynamic import() — loads module at runtime
src/index.ts · prose · downgraded · import('
Popular HTTP library — network access
src/classifier/dimensions.ts · prose · downgraded · got
Why the tier is capped
Execution sink present in raw bytes (Hard Floor: class A). Final tier capped at Caution — cannot be lifted by any downgrade, example-payload opt-in, or allowlist.
Permissions & capabilities
Requires 2 environment variables. (2 sensitive: ANTHROPIC_API_KEY, GOOGLE_AI_API_KEY). Requires 1 system binary.
Is this flag fair?
Thanks — recorded.